[Intel-gfx] Problems configuring unique xorg.conf file

[Alan W. Irwin]

On 2010-08-26 17:37-0000 Nasa wrote:

>
> ----- "Alan W. Irwin" <irwin at beluga.phys.uvic.ca> wrote:
>
>> On 2010-08-26 10:47-0000 Nasa wrote:
>>
>>> Hi,
>>>
>>>
>>> I am trying to configure a rather obscure resolution (800x480) which
>> is the
>>> native resolution for my monitor.  The monitor doesn't provide EDID
>> or DDC
>>> information (it's connected over VGA).  And the default settings by
>> the
>>> driver produces displays that don't fit the monitor properly.  I
>> would like
>>> to construct a xorg.conf file with appropraite vertical refresh
>> rates, horizontal
>>> syncs, and/or modelines to work correctly with the monitor.
>> However, there
>>> doesn't seem a way to turn off the driver defaults for those items.
>> I tried options
>>> like NoDDC, UseEDID, etc with no luck.  I also tried using xrandr to
>> change
>>> resolutions after X has started.  The results end up being worse
>> than the
>>> intial problem I was trying to fix (ie: the screen is bigger than
>> the area
>>> available to display it).  I expect my inability to find a suitable
>> solution
>>> is due to my lack of knowledge -- so I hope someone can fill me in
>> to what
>>> I am missing.  Thanks in advance,
>>
>> Earlier this year before I replaced my long-time Sony monitor with an
>> LCD, and upgraded from Debian Lenny to Debian testing, the results of
>> gtf and PreferredMode worked for me. For example, my xorg.conf file
>> for that monitor had the following lines in the Monitor section
>>
>> #gtf 1024 768 85
>> # 1024x768 @ 85.00 Hz (GTF) hsync: 68.60 kHz; pclk: 94.39 MHz
>> Modeline "1024x768_85.00"  94.39  1024 1088 1200 1376  768 769 772
>> 807
>> -HSync +Vsync
>> Option "PreferredMode" "1024x768_85.00"
>>
>> Of course, intead of using the above example, you will want to run
>> something like
>>
>> gtf 800 400 85
>>
>> from the command line (man gtf), paste the results to your Monitor
>> section, and consistently update the identification of the mode used
>> by PreferredMode.
>>
>> I emphasize the above configuration lines worked for an old version of
>> the
>> Intel driver (Debian Lenny), and I don't know whether they  would
>> work
>> for a modern version.  But it is worth a try.
>>
>> Alan
>
> Thanks Alan,
>
> I actually attempted this via CVT which ended up with horizontal sync out
> of range errors.  Reading the MAN page for CVT didn't show any options to
> put in options to set that.  Does GTF have this capability?

No, but it doesn't matter.  Play with either cvt of gtf (I don't think
there is any real difference between them) with a fixed resolution and
varying vertical refresh rate, and you will see that the horizontal
sync frequency of the generated mode line is proportional to the
vertical refresh you specify.  So for your desired resolution if the
generated horizontal sync is lower than your allowed range, increase
the vertical refresh until you have a value within the allowed range.
Or if it is above (extremely unlikely for such a low resolution)
reduce the vertical refresh.

I have assumed above that the horizontal frequency limits have been
set correctly for your particular monitor.  That is not always the
case.  Check your monitor manual for the correct vertical and
horizontal frequency limits and if your X log shows those are not
being discovered properly by X than specify the correct ranges using the
VertRefresh and HorizSync values in the Monitor section.  In my case
I used

HorizSync       30-96
VertRefresh     48-120

corresponding to values published in my Sony g200 manual, but your
monitor manual is very likely to require different ranges.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________